On 1/2/20 1:30 PM, Stefan Roese wrote:
Hi Mauro,
On 30.12.19 13:14, Mauro Condarelli wrote:
On 12/30/19 11:22 AM, Daniel Schwierzeck wrote:
Am 30.12.19 um 10:19 schrieb Mauro Condarelli:
I am having problems with this patch.
Problem is "reset"command fails (for my board) with:
=> reset resetting ... ### ERROR ### Please RESET the board ###
I traced down problem to "drivers/sysreset/sysreset-uclass.c" requesting "uclass_first_device(UCLASS_SYSRESET, &dev)", while "drivers/reset/reset-mips.c" defines:
static const struct udevice_id mtmips_reset_ids[] = { { .compatible = "mediatek,mtmips-reset" }, { } };
U_BOOT_DRIVER(mtmips_reset) = { .name = "mtmips-reset", .id = UCLASS_RESET,
... so UCLASS_SYSRESET list is empty.
What am I doing wrong? TiA! Mauro
do you have the according node with compatible string "mediatek,mtmips-reset" in your device-tree?
Yes, I do, but problem is elsewhere: "reset" command looks for a UCLASS_SYSRESET while drivers/reset/reset-mips.c (implementing "mediatek,mtmips-reset") provides a UCLASS_RESET.
I know too little about u-boot internals to understand which one I should "fix" (if any).
I just tested current mainline U-Boot (top-of-tree) on the GARDENA board and the reset works just fine:
=> reset resetting ...
U-Boot 2020.01-rc5-00042-g6cb87cbb14 (Jan 02 2020 - 13:27:15 +0100)
CPU: MT7628 Rev 1.2 - Boot from XTAL (3-Byte SPI Addr) Model: GARDENA smart Gateway (MT7688) DRAM: 128 MiB WDT: Started with servicing (60s timeout) Loading Environment from SPI Flash... SF: Detected XM25QH64A with page size 256 Bytes, erase size 4 KiB, total 8 MiB OK F-Data:factory-data version 1 detected Net: eth0: eth@10110000 Hit any key to stop autoboot: 0
I didn't check the details of this mail thread, so I can't really comment on this. Your board is so similar to the GARDENA one (LinkIt) that it should work there as well. Do you gave some differences in your defconfig?
My defconfig looks very different, mainly due to not having (currently) any network in u-boot, but I have USB strorage. I attach the whole vocore2_defconfig. To me it seems the only difference that *might* be relevant is I don't have CONFIG_SYSRESET_SYSCON=y. ... enable... compile... load... test... SUCCESS!! That was it. I was thinking reset should have been handled by "reset-mtmips.c" but apparently "mfd syscon reboot driver" is needed. THANKS. I'll send another iteration of my patches soon after a bit of testing.
Thanks, Stefan
Regards Mauro